Energy-saving Monitoring Technology Research For Coal-fired Power Plant Units

The overall power generation efficiency is not high in our country, there is plenty of potential to improve the utilization efficiency of energy, and the need of energy-saving monitoring improvement is growing fast. One of the problems is that lots of important technical parameters were difficult to be real-time measured online and many economic parameters as well. In the steam system aspect of this paper, Changshu power plant 1000 MW unit was carried out as the research object for energy-saving monitoring calculation and analysis, using the method of each extraction steam equivalent enthalpy drop based on thermal equilibrium analysis, general matrix equation form of the power plant thermal system was confirmed to calculate the enthalpy drop and the extraction efficiency of each extraction steam. The overall of the system was quantitatively analyzed as well as some certain parts, and where to do the improvements could be confirmed. In the aspect of the boiler system, the state grid energy Shentou second power plant 500 MW unit was carried out as the research object, soft sensor method was introduced to realtime measure the parameters which straightly reflects the boiler efficiency, such as oxygen content in flue gas, ball mill load and so on. Soft sensor method is to utilize some measurable parameters as the auxiliary variable to estimate the unmeasurable variable through on line analysis and calculation. This article analyzed the characteristics of the monitoring objects of coalfired power plants and found out the key factors that affect the accuracy of soft sensor model, which was the original data pre-processing, the optimizing selection of auxiliary variables, the calculation method of the model and so on. Then based on the original data pre-processing and the optimizing selection of auxiliary variables, this article introduced the modeling method of parameter self-adaptive support vector machine to reduce the human impacts and to improve the accuracy of the model. And the paper applied this method in the oxygen content in flue gas soft sensor modeling with the field data and successfully proved that the method was effective. The effective results of all the parameters above can be used by the power plant control system to optimize the operation of the power plant.